(Evidently forgot to create a tag for this earlier). Corresponds to PR https://github.com/nitrogenlab/pyompa/pull/4; description of changes reproduced here.
- Added support for "depth_default" in the penalty functions
- Rows with na values in the parameter columns for the observations are automatically dropped
- If an endmember name for which a usage penalty is specified is not in the endmembers df, an error now gets thrown